Pro/Forums - View Single Post

Cova · 05-20-2002, 03:33 PM

Ok - I don't often surf around the msg-boards much on the weekends (need something to do at work on mondays), but since monday is a holiday I'll make an exception and actually post this from the system pictured above

Answering questions from top to bottom

I'm just using the in-socket temp probe on the KT7A-RAID for measurements right now (reading it with MBM5.1 or so), and under 100% load right now (as I write this) the temp is 46C. I doubt it will go up any higher than that unless my ambient temp increases.

Cleanliness? I left the digi-cam at work but I'll get a picture from the angle of hte second pic of the top of the case, I just moved my "rats nest" of cabling up there (where I don't have any windows or plans to put any)

Ya - that sticker kicks ass, especially since it's an authentic one I stole from an OLD IMB mainframe computer. That IBM used to take up 3/4 of our computer room at work and we recently disposoed of it, so I have a few communication cards, a couple CPU's, that sticker, and prolly some other stuff around from it.

I haven't really gotten into the serious tweaking much (just finished re-installing windows yesterday), but it's currently running at 1533 at the max voltage that the KT7A will allow, and is more stable than it ever was with the HSF on it. The old HSF was a Volcano II and I couldn't go from 1333 to 1400 on that thing without problems. I don't know if I actually need the max voltage to stay at 1533 or not, but the water can take care of the extra heat and I'll prolly end up needing that voltage soon as I get into more tweaking.

I don't have a flow meter on it, so I don't know what the T-fitting does to the flow. I've also never run it without that fitting, so I can't compare temps with/without. But I know that it works really well as an air-trap forcing the water to turn like that instead of passing straight-through (straight-through may work, haven't tried that). I added a bit more water-wetter to my system after a couple days of running (my water was clear again, dunno where the water-wetter went), I just put a capful into the top of the fill-tube, and I could see the turbulence and mixing happening in the tygon just above the T. It looks like very little water actually does move up into the T, but the stuff that does slows down a lot while it churns around in there before heading back out, which is good because tiny air bubbles won't rise out of water that is moving too fast, they just get carried along with the flow.

Hmm...., temp reading is up to 47C now. Also, idle temp is usually around 33C if yer wondering.

Also, I made an interesting discovery the other day - not all 100% CPU usage loads create the same amount of heat. I've tried a bunch of different things lately that all pin the CPU at 100% usage (monitor that with XP Task Manager), but encoding a DVD into a DivX video is consistently 4 degrees hotter than anything else I've found so far. And it gets up to that temp quick, so it's not just cause it takes 10 hours to finish (and yes, thats what I have running in the background now thats making it hot - I consider my CPU stable when I can queue up 10 hours of work at 100% load and it finishes with no crashes and no video corruption in the resulting avi.)

05-20-2002, 03:33 PM	#8
Cova Cooling Savant Join Date: May 2002 Location: Canada Posts: 247	Ok - I don't often surf around the msg-boards much on the weekends (need something to do at work on mondays), but since monday is a holiday I'll make an exception and actually post this from the system pictured above Answering questions from top to bottom I'm just using the in-socket temp probe on the KT7A-RAID for measurements right now (reading it with MBM5.1 or so), and under 100% load right now (as I write this) the temp is 46C. I doubt it will go up any higher than that unless my ambient temp increases. Cleanliness? I left the digi-cam at work but I'll get a picture from the angle of hte second pic of the top of the case, I just moved my "rats nest" of cabling up there (where I don't have any windows or plans to put any) Ya - that sticker kicks ass, especially since it's an authentic one I stole from an OLD IMB mainframe computer. That IBM used to take up 3/4 of our computer room at work and we recently disposoed of it, so I have a few communication cards, a couple CPU's, that sticker, and prolly some other stuff around from it. I haven't really gotten into the serious tweaking much (just finished re-installing windows yesterday), but it's currently running at 1533 at the max voltage that the KT7A will allow, and is more stable than it ever was with the HSF on it. The old HSF was a Volcano II and I couldn't go from 1333 to 1400 on that thing without problems. I don't know if I actually need the max voltage to stay at 1533 or not, but the water can take care of the extra heat and I'll prolly end up needing that voltage soon as I get into more tweaking. I don't have a flow meter on it, so I don't know what the T-fitting does to the flow. I've also never run it without that fitting, so I can't compare temps with/without. But I know that it works really well as an air-trap forcing the water to turn like that instead of passing straight-through (straight-through may work, haven't tried that). I added a bit more water-wetter to my system after a couple days of running (my water was clear again, dunno where the water-wetter went), I just put a capful into the top of the fill-tube, and I could see the turbulence and mixing happening in the tygon just above the T. It looks like very little water actually does move up into the T, but the stuff that does slows down a lot while it churns around in there before heading back out, which is good because tiny air bubbles won't rise out of water that is moving too fast, they just get carried along with the flow. Hmm...., temp reading is up to 47C now. Also, idle temp is usually around 33C if yer wondering. Also, I made an interesting discovery the other day - not all 100% CPU usage loads create the same amount of heat. I've tried a bunch of different things lately that all pin the CPU at 100% usage (monitor that with XP Task Manager), but encoding a DVD into a DivX video is consistently 4 degrees hotter than anything else I've found so far. And it gets up to that temp quick, so it's not just cause it takes 10 hours to finish (and yes, thats what I have running in the background now thats making it hot - I consider my CPU stable when I can queue up 10 hours of work at 100% load and it finishes with no crashes and no video corruption in the resulting avi.)